Study Of Symbol Estimation Algorithm For Underwater Acoustic Communication System With Multiplicative Noise

One of the important subjects of modern signal processing is that the useful signalshould be estimated as accurately as possible from the observation data mixed with thenoise and interference in a certain sense through the optimal criterion. The multiplicativenoise system model is proposed in this paper to characterize the underwater acousticchannel. And the transmitted symbol is estimated based on the optimal filtering andsmoothing recursive algorithms in the linear minimum variance criterion.There are multi-paths effects in the underwater acoustic channels, following withfading, scattering, so the inter-symbol interference (ISI) is very serious in received thesignal. Besides, the channel loss characteristics vary over time, which makes the channelgain changed in a fluctuating state with normal distribution. Therefore, the multiplicativenoise is used to characterize the random fluctuation of the channels and the underwateracoustic channels can be modeled as the system with multiplicative noise. The mainstudy of the dissertation is as follows:1、For the single-input multiple-output system, based on the expression of thediscrete base band received signal,using the estimated underwater acoustic channelparameters, through the introduction of multiplicative noise, the state space model withmultiplicative noise is established. Then the underwater signal processing is transferredto the state-space model with multiplicative noise.2、The transmitted signal of the underwater acoustic communication system is basedon the binary phase shift keying (bpsk) or quadrature phase shift keying (qpsk) signal,which is binary or four-value and obey a certain probability distribution. So the optimalfiltering recursive algorithm is specifically developed for the state space model withnongaussian dynamic noise in the sense of linear minimum-variance based on theprojection theorem, and then the optimal smoothing recursive algorithm is obtained.3、Because of the correspondence between the state’s first dimension and thetransmitted symbol signal, the filtering and smoothing algorithms are used to estimatethe transmitted symbol.4、Simulation results show that estimation of the underwater symbol signal based onfiltering and smoothing algorithms can effectively overcome the ISI and noise pollution,which validates the algorithms.